Using Machine Learning to Predict Abnormal Carotid Intima-Media Thickness in Type 2 Diabetes

Area of Science:

• Cardiology
• Medical Informatics
• Diabetes Research

Background:

• Carotid intima-media thickness (c-IMT) is a validated marker for cardiovascular disease (CVD) risk, particularly in patients with type 2 diabetes (T2D).
• Accurate prediction of c-IMT is crucial for early CVD risk stratification in T2D cohorts.

Purpose of the Study:

• To compare the predictive performance of various machine learning (ML) algorithms against traditional logistic regression for c-IMT in T2D patients.
• To identify the most significant baseline risk factors associated with c-IMT progression in this cohort.

Main Methods:

• A cohort of 924 T2D patients was followed for four years.
• Machine learning models (classification and regression tree, random forest, eXtreme gradient boosting, Naïve Bayes) and multiple logistic regression were employed to predict c-IMT.
• Model performance was evaluated using the area under the receiver operation curve (AUC).

Main Results:

• Most ML methods demonstrated comparable or superior performance to logistic regression in predicting c-IMT, as indicated by higher AUC values.
• Classification and regression tree was the only ML method not outperforming logistic regression.
• Significant predictors of c-IMT included age, sex, creatinine, body mass index, diastolic blood pressure, and diabetes duration.

Conclusions:

• Machine learning approaches offer enhanced prediction capabilities for c-IMT in T2D patients compared to conventional logistic regression.
• These findings support the integration of ML for improved early identification and management of cardiovascular risk in type 2 diabetes.